The 1st Workshop on Reimagining Distributed Computing for LLMs (DistLLM)
aims to tackle the pressing challenges in the scalable, efficient, and
reliable distributed training, fine-tuning, and inference of large language
models (LLMs) and deep neural networks (DNNs). This workshop creates a
platform to bring together leading experts in distributed systems,
networking, and AI to collaboratively explore innovative solutions across
multiple critical domains. These include advanced multi-dimensional
parallelism techniques encompassing data, tensor, pipeline, and expert
parallelism; communication and synchronization protocols optimized for
efficient collective communication and congestion control; and intelligent
distributed scheduling strategies that address multi-tenant resource
allocation and adaptive scheduling under communication bottlenecks.
Further, the workshop emphasizes network-aware system design to ensure
load-balanced traffic and topology-aware flow control, alongside developing
resilience and fault tolerance methods such as checkpointing and anomaly
detection for long-running training jobs. Recognizing the growing
complexity of LLM training infrastructure, the
event also addresses emerging challenges of multi-cluster training spanning
data centers and edge clusters, including inter-domain and wide area
network scheduling. Additionally, specialized topics on LLM deployment
tailored for edge computing environments—such as fine-tuning tiny LLMs and
optimizing for heterogeneous compute devices—are featured. Through a
comprehensive program of refereed papers, keynote talks, expert panels,
lightning talks, and
demonstrations, the DistLLM workshop seeks to advance the state-of-the-art
in distributed systems infrastructure for AI while leveraging LLM
capabilities to innovate distributed computing itself. This workshop
intends to drive transformative research that meets the extraordinary
demands of modern AI workloads and bridges key gaps in scalable, efficient,
and resilient distributed model training and deployment.
Topics of interest include but are not limited to:
1. Parallelism-aware scheduling and job orchestration
2. Scalable parallel algorithms for data, model, pipeline, and expert
parallelism
3. Collective protocols and congestion mitigation
4. Fault tolerance, checkpointing, and recovery for long-running training
5. Multi-tenant cluster scheduling and resource sharing
6. Geo-distributed and edge-coordinated model training
7. System design for MoE, long-sequence models, and sequence parallelism
8. Tools and frameworks for visualizing, debugging, or optimizing training
systems
9. Security, Privacy, and Isolation-aware distributed training
10. Protocols for consistent and efficient tracking of model checkpoints
11. Algorithms for fair and concurrent training of multiple models
12. LLMs for edge compute nodes
Important Dates: (All Dates are Anywhere on Earth)
● Submission deadline: September 15th, 2025
● Notification to authors: October 25th, 2025
● Camera-ready paper due: November 15th, 2025
● Workshop date: January 6-9th, 2026 (Nara, Japan) [Exact date to be
announced] Paper
Submission Guidelines:
All papers must be original and not simultaneously submitted to another
journal or conference. All papers will bepeer-reviewed using a double-blind
peer-review process by at least three members of the program committee.
Submissions should be a complete manuscript. DistLLM accepts Full papers: 6
pages in ACM Conference format
(including title, abstract, figures, and references).
Workshop Paper Format Guidelines (ACM Style Adaptation):
● Papers should be formatted in double-column, single-spaced layout using a
10-point font size on standard 8.5 x 11-inch (US letter) pages.
● Submissions are anonymous. The conference will employ a lightweight
double-blind reviewing process. Manuscripts should not include author names
and affiliations.
● Submissions should not reveal the identity of the authors in any way.
Authors should ensure that any references to their own related work are in
the third person (e.g., not "We build on our previous work ..." but rather
"We build on the work of ...").
● Authors are required to use the official ACM conference templates for
manuscript preparation, available in both MS Word and LaTeX formats.
● ACM templates ensure compliance with ACM's publication standards and can
be found here:
https://www.acm.org/publications/proceedings-template
● Papers are to be submitted electronically in PDF format. Submitted papers
should not have appeared in or be under consideration for a different
workshop, conference or journal. All accepted papers need to be presented
at the workshop by one of the authors.
● All accepted papers (subject to post-review revisions) will be published
in the ICDCN 2026 companion proceedings.
● Submission Link: https://easychair.org/conferences?conf=icdcn2026, Select
WS2: 1st Workshop on Reimagining Distributed Computing for LLMs for
submissions.
Here are the guidelines for Artificial Intelligence (AI)-Generated Text and
related policies for the workshop papers, based on the provided ACM
policies and the instructions given:
Authorship and Use of Generative AI:
Ensure that the ACM Policy on Authorship is followed strictly for all
accepted papers. This means all authors must be identifiable human beings
who made substantial intellectual contributions and take responsibility for
the work. Generative AI tools and technologies such as ChatGPT may be used
to assist in creating sections of the work (text,
code, data, citations, etc.), but these must be fully disclosed in the
acknowledgements section of the paper. Basic word processing assistance
(spell check, grammar correction) does not require disclosure.
Authorship cannot be added or removed after paper acceptance.
Open Publication Model and Article Processing Charges (APC):
ICDCN 2026 workshop proceedings will be published as a companion volume
along with the main conference proceedings. However, please note that ACM
has moved to a new open-access publishing model for all conference
proceedings to be published via ACM ICPS. The authors have to pay an
Article Processing Charge (APC) to ACM (which is beyond the regular
conference registration fee) if the corresponding author's organization is
not a member
of the ACM Open program. Please check the details through this link:
https://www.acm.org/publications/icps/author-guidance.
Several institutes worldwide are already members of the ACM Open program.
The authors can use the following link to check if their organization is a
member under the
ACM Open program: https://libraries.acm.org/acmopen/open-participants. For
any clarifications, contact icps-info@acm.org.
Here are the specific ACM links related to the policies and guidelines
mentioned for the workshop papers:
ACM Policy on Authorship:
https://www.acm.org/publications/policies/new-acm-policy-on-authorship
ACM Policy on the use of Generative AI in papers (FAQ):
https://www.acm.org/publications/policies/frequently-asked-questions
ACM Open Publication Model and FAQs for International Conference
Proceedings Series (ICPS): https://www.acm.org/publications/icps/faq
List of Institutions Participating in the ACM Open Program:
https://libraries.acm.org/acmopen/open-participants
Guidance on Article Processing Charges (APC) and publishing in the ACM
Digital Library:
https://www.acm.org/publications/icps/author-guidance
ICDCN 2026 Main Conference Paper Submission Guidelines (including
double-blind review process):
https://sites.google.com/view/icdcn2026/submissions/call-for-paper?authuser=0
These links will help authors and organizers ensure compliance with ACM's
policies on authorship, generative AI usage, open access publishing, and
paper submission guidelines.
Regards,
Subhrangsu Mandal
**********************************************************
*
* Contributions to be spread via DMANET are submitted to
*
* DMANET@zpr.uni-koeln.de
*
* Replies to a message carried on DMANET should NOT be
* addressed to DMANET but to the original sender. The
* original sender, however, is invited to prepare an
* update of the replies received and to communicate it
* via DMANET.
*
* DISCRETE MATHEMATICS AND ALGORITHMS NETWORK (DMANET)
* http://www.zaik.uni-koeln.de/AFS/publications/dmanet/
*
**********************************************************
Posts
